Spatial Competence Benchmark

Computer Science > Artificial Intelligence [Submitted on 5 Mar 2026] Spatial Competence Benchmark Jash Vira, Ashley Harris Spatial competence is the quality of maintaining a consistent internal representation of an environment and using it to infer discrete structure and plan actions under constraints. Prevailing spatial evaluations for large models are limited to probing isolated primitives through 3D transformations or visual question answering. We introduce the Spatial Competence Benchmark (SCBench), spanning three hierarchical capability buckets whose tasks require executable outputs verified by deterministic checkers or simulator-based evaluators. On SCBench, three frontier models exhibit monotonically decreasing accuracy up the capability ladder. Sweeping output-token caps shows that accuracy gains concentrate at low budgets and saturate quickly, and failures are dominated by locally plausible geometry that breaks global constraints. We release the task generators, verifiers, and visualisation tooling. Comments: Accepted at the ICLR 2026 Workshop on Efficient Spatial Reasoning Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2604.09594 [cs.AI]   (or arXiv:2604.09594v1 [cs.AI] for this version)   https://doi.org/10.48550/arXiv.2604.09594 Focus to learn more Submission history From: Jash Vira [view email] [v1] Thu, 5 Mar 2026 00:27:55 UTC (3,304 KB) Access Paper: HTML (experimental) view license Current browse context: cs.AI < prev   |   next > new | recent | 2026-04 Change to browse by: cs cs.LG References & Citations NASA ADS Google Scholar Semantic Scholar Export BibTeX Citation Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Demos Related Papers About arXivLabs Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)